cartesian_prod

→Tensor

cartesian_prod(tensors: Tensor = ())

source

Return the Cartesian product of the given 1-D tensors.

Generalisation of nested itertools.product to tensor inputs. Given k tensors of sizes n_1, ..., n_k the output has shape (prod(n_i), k).

Parameters

*tensorsTensor= ()

Any number of 1-D tensors.

Returns

Tensor

2-D tensor whose rows are all combinations of input elements, in lexicographic order.

Notes

For 2-D-style grid coordinates (separate per-axis outputs) use meshgrid with indexing='ij'.

Examples

>>> import lucid
>>> a = lucid.tensor([1, 2])
>>> b = lucid.tensor([3, 4])
>>> lucid.cartesian_prod(a, b)
Tensor([[1, 3],
        [1, 4],
        [2, 3],
        [2, 4]])